One of the requirements currently imposed (by the FCC) upon communication systems and networks is that they comply with very stringent spurious energy suppression standards. In today's digital communications environment, data signals (which have square rather than sinusoidal characteristics) are typically conditioned and modulated upon a higher carrier frequency (e.g., via emitter coupled logic circuitry (ECL)-based in-phase I and quadrature Q channels) where the data can be more readily transmitted.
Since the primary signal energy component of interest is located in the spectrum immediately surrounding the carrier center frequency, other spurious energy--typically the odd harmonics (multiples) of the original carrier frequency resulting from the modulation (multiplication or mixing) operation of such non-sinusoidal signals--must be removed, in order to avoid contaminating adjacent carrier channels that contain their own information energy components.
One way to remove such spurious energy is to apply the mixer output to a lowpass filter having a cutoff frequency that is slightly higher than the highest carrier frequency to be employed in a given communication system, such as a frequency-agile transceiver. Unfortunately, if the highest carrier frequency capable of being generated is close to the second or third harmonic of the lowest carrier frequency in the operating range of the system, the size (order) of the filter required to achieve sufficient unwanted energy suppression under all operating conditions is unacceptably large.
As a non-limiting example, in the case of conducting data communications over a television cable network, the required harmonic suppression filter may be an eight to twelve order filter. Such a large and complex filter is obviously impractical where cost and semiconductor silicon area are prime considerations in chip design. Moreover, such a large order filter has a higher group delay error, which can increase bit error rate.
Another technique is to add a filter pole to the carrier generator prior to the mixer. For a fixed carrier, adding a single pole is readily accomplished by simply adding a capacitor; however, in the case of a frequency-agile system, it is necessary to employ a plurality of capacitors, typically implemented as a programmable (controllably switched, binary weighted) capacitor array.
If the process employed to manufacture the integrated circuit architecture is a BiMOS process, MOSFET switches may be used to switch among the capacitors of the array. However, for a multiple channel system where there are a considerable number of potential carrier frequencies--requiring a separate capacitor for each carrier frequency, a binary-weighted capacitor array cannot be effectively employed, since the filter's cutoff frequency varies with the reciprocal of the value of the capacitor. As a consequence, if the cutoff frequency is to vary in a linear manner, the cutoff control capacitor must vary in a non-linear fashion, which not only again leads to an unacceptably large and complex filter, but the required switch array introduces parasitic poles, which are not readily compensated.